Beam report compression based on linear principal component analysis
By using a linear principal component analysis filter to compress beam reports in wireless communication, the problems of high signaling overhead and computational resource consumption are solved, resulting in interpretable output and improved network performance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- QUALCOMM INC
- Filing Date
- 2024-08-20
- Publication Date
- 2026-06-02
AI Technical Summary
In wireless communication, existing technologies suffer from high signaling overhead and computational resource consumption in beam reporting, especially when the uplink control information load is limited, and the compression method of the automatic encoder output is uninterpretable and computationally complex.
A method based on linear principal component analysis (PCA) filters is used to compress beam-related measurement information. This is achieved by generating interpretable maximum singular value outputs and discarding smaller singular value terms in the signaling to reduce signaling overhead. At the same time, the singular value decomposition pre-computed by network nodes is used to reduce the computational resource consumption of user equipment.
It achieves the goal of reducing signaling overhead and computational resource consumption while providing interpretable beam report output, supporting effective training and inference of AI/ML models, and improving network performance.
Smart Images
Figure FT_1 
Figure FT_2 
Figure FT_3